Wireless Fingerprint Reconstruction for Location Confirmation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless localization techniques face challenges in accurately confirming delivery locations due to incomplete wireless fingerprints, which are often incomplete due to hardware and software limitations in wireless scanning, leading to inaccurate model parameter estimation and reduced accuracy in distance metrics during the survey and online confirmation phases.
Innovation Solution
The proposed solution involves generating an attribute prediction model using multiple fingerprint samples to estimate missing attributes in wireless fingerprints, creating an imputed fingerprint that includes these estimated attributes, which is then used for more accurate wireless localization by comparing it against a location confirmation model to confirm the delivery location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If wireless scanning is performed using current hardware and software, then location confirmation can be achieved, but the wireless fingerprints are incomplete leading to reduced accuracy
Solution Approach 1:
The system performs preliminary wireless scanning during a survey phase to collect fingerprint samples from multiple locations. These samples are used to train machine learning models before the online phase, enabling the system to predict missing attributes and improve location confirmation accuracy even when individual scans are incomplete
Solution Approach 2:
Machine learning models serve as intermediaries between the incomplete wireless scan data and the location confirmation process. The models predict missing wireless attributes (such as signal strength from undetected access points) by learning patterns from training data, effectively bridging the information gap caused by hardware and software limitations
2Device complexity
If hardware and software limitations are accepted, then device complexity is reduced, but wireless fingerprint accuracy deteriorates
Solution Approach 1:
The system uses the data collected from incomplete scans themselves to train machine learning models that can predict the missing information. Rather than requiring more sophisticated hardware, the system leverages its own operational data to improve accuracy, with the models learning to infer undetected wireless attributes from the patterns in the collected samples
3Measurement precision
If multiple fingerprint samples are collected for training, then model accuracy is improved, but data collection time increases
Solution Approach 1:
The system collects more fingerprint samples than the minimum single scan would provide during the survey phase, using multiple samples from the same location to train more accurate machine learning models. This excessive data collection improves the models' ability to predict missing attributes, compensating for the additional time invested in the survey phase by enabling faster, more accurate location confirmation during online deliveries
Data Source
AI summary
Disclosed are various embodiments for estimating missing wireless attributes of a wireless fingerprint generated for a given location and confirming that the given location is a correct location. A collection of wireless fingerprint samples for the given location can be used to train an attribute prediction model that can estimate missing attributes in wireless fingerprints. For example, a wireless fingerprint may be incomplete as the fingerprint may fail to include attribute data for some AP devices associated with the given location. The attribute prediction model can be used to reconstruct wireless fingerprints to include the missing attributes. The reconstructed wireless fingerprints can then be used to train a location confirmation model which can be used for accurate location confirmation.